The United States and China are preparing to discuss the safety risks of artificial intelligence during a dialogue scheduled for mid-September, at a time when rapidly developing artificial intelligence capabilities are reaching a global turning point, according to Reuters, citing two informed sources.
The two sources said that the talks will be the first official bilateral discussions dedicated to artificial intelligence between the United States and China since President Donald Trump took office for a second term, and will be led from the American side by Treasury Secretary Scott Besent. However, a White House official said that “there is currently no meeting related to artificial intelligence scheduled for mid-September.”
Trump and Chinese President Xi Jinping are scheduled to meet on September 24 during a summit in Washington.
One of the sources said that the United States wants to discuss cooperation on monitoring cyber attacks directed by artificial intelligence, and has put forward a proposal asking American and Chinese artificial intelligence laboratories to “monitor themselves” and exchange information to prevent cyber attacks linked to artificial intelligence.
Last week, OpenAI and cybersecurity researchers said that nearly 700 out-of-control artificial intelligence agents built on OpenAI models hacked into the artificial intelligence startup Hacking Face in July and attempted to cover its tracks by falsifying records.
Reuters reported on Friday that a swarm of Open AI agents hacked a website in Germany this spring and turned it into a bulletin board for other artificial intelligence agents.
“The talks between the two AI superpowers come at the most crucial moment,” said Paul Triolo, a partner at DGA-Albright Stonebridge. “It is either now or never.”
Washington is also concerned about the possibility of the emergence in the future of a Chinese model of the “Methos” level, possessing the capabilities to carry out electronic attacks. The American technology company Anthropic is the developer of the Mythos model.
In June, White House Senior Advisor for Science and Technology Michael Kratsios accused the Chinese company Moonshot AI of stealing outputs from Anthropic's Fable model to help develop its version of K3, saying that the Chinese company was working to distill Anthropic's model.
Distillation refers to the process of training smaller AI models using the output of larger, more expensive models, with the goal of reducing the costs of training a new, powerful AI tool.
Fitch Ratings Agency removed Qatar from the “negative watch” list and kept its sovereign rating at “AA,” indicating a decline in the risks threatening liquefied natural gas facilities since March.
But the agency kept Qatar's outlook at "negative", in light of the continuing risks resulting from the disruption of liquefied natural gas exports through the Strait of Hormuz.
Qatar is among the economies that witnessed the largest reductions in expectations during a Reuters poll, as the gross domestic product is expected to contract by 8.1 percent this year, compared to expectations of a 6 percent decline last April.
The agency said in a statement on Friday evening: “It will take longer to absorb the impact of the war on creditworthiness.”
Qatar, one of the world's largest exporters of liquefied natural gas, has been dealing with disruptions in exports through the Strait of Hormuz and shortages linked to damage to energy facilities.
The rating confirmation comes months after Fitch warned against downgrading Qatar's sovereign rating, with mounting concerns about the security and economic repercussions of the war on Iran.
Reuters reported last month that Qatar has lost about $24 billion in gas sales since the conflict began six months ago.
Standard & Poor's and Moody's also confirmed Qatar's rating earlier this year, saying that its "substantial" financial position mitigates the impact of the US-Israeli war on Iran.
The British Treasury said it would lower a key rate used to evaluate the long-term benefits of investing in public infrastructure projects, as part of plans to boost investment across the country.
The discount rate used to evaluate public spending will be reduced to 3 percent from 3.5 percent, the Treasury Department said in a statement on Friday evening, clarifying amendments to the “Green Book” guide that regulates the approval of capital projects.
Discounting refers to the way the Treasury compares costs and benefits at different times, with £100 today being treated as more valuable than a promise to receive £100 in the future.
Lowering the discount rate narrows this gap, making it easier for long-term projects to show their full value rather than diminishing their value simply because their benefits take years to realize.
The government will publish full details of the plan, including its response to the discount rate review, in the Budget on October 28. Treasury Secretary John Healey plans to deliver his first major speech in his new role on Monday, where he is expected to outline details of the plan.
The Treasury said the changes aim to give transport, housing and social infrastructure projects “fairer treatment” in government spending decisions, by giving greater weight to benefits that take years to appear.
